When and where was BMLL founded?

BMLL was founded in 2014 in the machine learning laboratories of the University of Cambridge. BMLL launched BMLL Data Lab in 2017, BMLL Data Feed in 2019 and BMLL Vantage in 2022. BMLL started with London Stock Exchange Group, Borsa Italiana, CME, Euronext and Tradeweb data in 2015. It added US markets in 2018, Level 3 futures in 2021, and markets in Asia-Pacific, Canada, Brazil, Israel, Poland, South Africa, Turkey and India between 2023 and 2024. Options data followed in 2024. Today BMLL covers more than 145+ venues across equities, ETFs, futures, options and prediction markets. 

What is BMLL?

BMLL Technologies (BMLL) is an independent provider of harmonised Level 3, Level 2 and Level 1 historical market data and analytics. BMLL covers more than 140 trading venues across global equities, ETFs, futures, options and prediction markets, and delivers its data T+1. BMLL collects raw market data from exchange matching engines, co-located capture partners and, for some venues, data aggregators. It then converts that data into a single data model that works the same way across every venue. BMLL’s ready-made analytics are built on the same harmonised data, so the raw data and the analytics always match. BMLL is used by systematic trading firms, banks, brokers, asset managers, exchanges and regulators. They use it for execution analytics, market structure research, generating trading ideas, systematic trading, AI and machine learning training, surveillance and policy analysis. 

Who owns BMLL?

In October 2025, Nordic Capital acquired BMLL in close partnership with BMLL’s management team. Optiver, the global market maker, remains a minority shareholder. Before the acquisition, BMLL raised a $21 million strategic investment led by Optiver in October 2024, $26 million in Series B funding in 2022 and 2023 (which Snowflake Ventures joined), and $36 million in Series A and seed funding. 
 

Is BMLL an independent data provider?

Yes. BMLL is an independent, neutral third-party provider of historical market data and analytics. BMLL is not an exchange or a broker, so its data is used to compare venues, brokers and execution quality from a neutral standpoint.  Independence is why exchanges, brokers, asset managers and regulators all use BMLL to benchmark execution and market quality. Exchanges use BMLL to show members independent evidence of their venue’s quality. Asset managers use BMLL to assess brokers. Regulators use BMLL to check market quality without relying on data supplied by the firms they regulate. 
 

Who uses BMLL?

BMLL is used by systematic trading firms, asset managers, commodity trading advisers (CTAs), investment banks and brokers, exchanges and trading venues, ETF issuers, independent software vendors and regulators. BMLL is also used by academic institutions. Within those organisations, BMLL data is used by quants, market microstructure researchers, execution analysts, smart order routing and algorithm teams, surveillance teams and policy analysts. Firms that use or partner with BMLL include Optiver, Kepler Cheuvreux, Jefferies, Berenberg, SIX, Aquis Exchange, Exegy, Broadridge and Snowflake. 

How is BMLL different from other historical market data providers?

Where Level 3 source data is available, BMLL builds its datasets from it, ideally from raw packet capture, and converts it into one lossless, consistent global format with nanosecond timestamps. Where it isn’t, BMLL works from the most granular source available for that venue. BMLL is independent, holds the exchange licences as Vendor of Record. Because BMLL Level 1 and Level 2 data come from Level 3 data rather than from aggregated feeds, they are more accurate, consistent and transparent. BMLL’s ready-made analytics are built on the same data, so the raw data and the analytics always match. 
 

Is BMLL secure?

Yes. BMLL is certified to ISO 27001, the international standard for information security management. BMLL Data Lab includes enterprise-grade security features, including single sign-on (SSO) and version control. 

What is Level 3 market data?

Level 3 market data is the full-depth order book at the level of individual messages. Every order insert, update, cancellation and execution is captured. BMLL Level 3 data keeps matching-engine timestamps to the nanosecond and exchange sequence numbers for exact replay. With Level 3 data, the full lifecycle of every order can be rebuilt from the message stream and joined by order ID. That makes it possible to see each order’s position in the queue, how long it rested at a price, its chance of being filled, how it was modified and what happened after it was cancelled. BMLL also provides inferred orders. BMLL builds its Level 1 and Level 2 data directly from Level 3 data, so all three levels are consistent with each other. 
 

What is Level 2 market data?

Level 2 market data shows full order book depth added up by price level, with the total volume and number of orders at every level. BMLL builds its Level 2 data directly from Level 3 source data. Level 2 shows how much liquidity is available at each price, but not the individual orders behind it. Use Level 3 when you need queue position, order lifecycles or fill analysis. 
 

What is Level 1 market data?

Level 1 market data shows only the best bid and best ask (top of book), with total volume and number of orders at each. BMLL builds its Level 1 data directly from Level 3 source data. Level 1 data is enough for simple price and spread analysis. It cannot show depth beyond the best prices or the behaviour of individual orders, which is why BMLL recommends Level 3 for TCA, surveillance and backtesting. 
 

What is the difference between Level 1, Level 2 and Level 3 data?

Level 1 data shows only the best bid and ask. Level 2 data shows total depth at every price level. Level 3 data shows every individual order and every message that affects it. BMLL provides all three, and builds Level 1 and Level 2 from Level 3. Where Level 3 source data is available, all three levels come from the same source, so BMLL Level 1, 2 and 3 data match exactly. Traditional providers often build Level 1 and Level 2 from aggregated feeds, which can lead to gaps and mismatches. 
 

What is the difference between harmonised and normalised data?

BMLL harmonised data is lossless: every field specific to a venue is kept, and the data model is the same across venues. BMLL normalised data is a smaller, consistent set of fields for each venue, built from BMLL’s global data model. Harmonised data means the order book rebuilds and replays the same way on every venue. Normalised data makes cross-venue analysis simpler. For example, BMLL’s normalised schema has five auction types, where some other providers carry 31 or more. 

How does BMLL source its market data?

BMLL works from the most granular source available for every venue it covers. Wherever possible, that is raw packet capture (PCAP) taken directly from the exchange or in colocation. Where PCAP is not available, BMLL uses the closest available source and applies the same engineering standards. BMLL keeps nanosecond timestamps, market state codes, trade condition codes and reference data from the source. BMLL collects both A/B feeds and snapshot feeds. BMLL keeps the original source data indefinitely, so it can always regenerate the entire history from source when exchange specifications change. 

Which BMLL product should I choose?

Choose BMLL Data Lab if you want to write Python code on BMLL data in a ready-made environment. Choose BMLL Data Feed if you want BMLL data delivered into your own systems. Choose BMLL Vantage if you want ready-made, no-code analytics. All three BMLL products use the same harmonised BMLL data model, and many clients use more than one. For example, a team might use BMLL Vantage for quick venue comparisons and BMLL Data Lab for deeper research. 

What is BMLL Data Lab?

BMLL Data Lab is a cloud-based Python data science platform for historical market data. It combines BMLL’s full harmonised Level 3 data for equities, futures, options and prediction markets with built-in analytics and computing power that scales across CPUs and GPUs. BMLL Data Lab is fully managed and can process datasets of many petabytes without infrastructure bottlenecks. It includes easy-to-use APIs, built-in analytics, version-controlled notebooks, single sign-on (SSO) and, in Data Lab Enterprise, scheduled jobs and connections to your own cloud storage and SFTP servers. 
BMLL Data Lab is used by quant researchers, execution and TCA teams, market structure researchers, smart order routing and algorithm teams, and teams assessing new markets. 

What is the difference between the BMLL Data Lab trial and the Enterprise version?

The BMLL Data Lab trial gives you instances of up to 16 GB and access to most datasets, so you can explore the environment. BMLL Data Lab Enterprise offers instances of up to 1.5 TB, a custom choice of datasets, scheduled jobs, shared workspaces, large clusters, GPUs, cloud storage connections and dedicated support. BMLL Data Lab Enterprise can run large Elastic MapReduce (EMR) clusters with 64 cores per node for Apache Spark processing. It connects to Amazon S3, Azure Blob Storage and Google Cloud Storage. Jobs can be scheduled by market, by time or triggered through the API. Some markets are not available in the BMLL Data Lab trial because of exchange licensing. 

What is BMLL Data Feed?

BMLL Data Feed delivers petabytes of BMLL historical market data and analytics directly into your own environment. Options include API, files, a cloud data warehouse or a data platform. The data has Level 3 order book detail, is normalised across venues and regions, and arrives in a research-ready format. BMLL Data Feed suits firms that want to run BMLL data inside their own systems, such as Snowflake, Databricks, AWS, Azure or Google Cloud, alongside their own trading records. 
 

What is BMLL Vantage?

BMLL Vantage is a no-code visual analytics app for comparing liquidity, venue quality and market fragmentation across global equities and ETFs. BMLL Vantage provides more than 500 metrics drawn from BMLL Level 3 data. BMLL Vantage lets trading, execution, sales, strategy and regulatory teams analyse single stocks, custom portfolios or entire markets without writing code. Users can share and comment on charts and export charts and data to Excel. BMLL Vantage also works with BMLL Data Lab and BMLL Data Feed for more detailed custom analysis. Common uses of BMLL Vantage include comparing liquidity across exchanges, alternative trading venues and trade reporting facilities. It is also used to analyse closing auctions and index rebalances, to prove a venue’s quality, and to assess new markets. 
 

What is BMLL Trades Plus?

BMLL Trades Plus adds detail to every trade at the level of individual messages. Each trade carries MMT flags, BMLL’s 22-value liquidity classification, block and participant type, and the best bid and ask at the time of execution. It also includes mid-prices before and after the trade at 15 intervals from 1 millisecond to 5 minutes. Participant types in BMLL Trades Plus are retail, institutional, market maker and broker-dealer. BMLL Trades Plus also includes the best bid and offer after the trade, measured at 1, 3 and 10 times the trade size. BMLL Trades Plus is built for TCA, broker benchmarking, markout analysis and execution governance. It removes the repetitive work of joining trade data with quote data. 
 

What is BMLL Daily Classified Trades?

BMLL Daily Classified Trades adds up trades each day by listing, MIC (market identifier code), execution venue and BMLL classification. It shows volume, number of trades and value traded in local currency, US dollars and euros. BMLL’s 22-value classification covers categories such as lit on-book trading, dark trading below the large-in-scale threshold, periodic auctions, OTC trades, systematic internalisers, off-book trades on exchange and benchmark trades. It lets users separate liquidity they can actually access from liquidity they cannot, across a whole portfolio. 
For US equities, BMLL includes FINRA OTC Transparency data weekly, going back to December 2018. 
 

What is BMLL Daily Analytics?

BMLL Daily Analytics is a library of ready-made daily metrics that can be queried through time_series.query(). Metrics include spread, depth, liquidity around the best bid and offer, venue price quality, fill probability, liquidation cost, market impact and auction analysis. BMLL Daily Analytics also covers volatility, OHLC, TWAP, VWAP, retail flow, quote-to-trade and trade-to-order ratios, average resting time and volume at the best price. Venue price quality is measured by how often a venue sets, joins or lags the European best bid and offer (EBBO). BMLL documents each metric, including how it is built, the time intervals it uses and what it depends on. 
 

Is BMLL cheaper than building a historical market data platform in-house?

BMLL’s cloud-based data infrastructure, scalable computing and data warehouse are licensed for a fraction of the cost of building them in-house. BMLL also manages data stewardship, infrastructure and computing, which further lowers the total cost of ownership. With BMLL, users spend their time on work that adds value, rather than the 80% of time often spent sourcing, cleaning and preparing data. 
 

Which markets does BMLL cover?

BMLL covers more than 140 trading venues across the Americas, Europe, the Middle East and Africa, and Asia-Pacific. BMLL coverage includes primary exchanges, multilateral trading facilities (MTFs), alternative trading systems (ATSs) and trade reporting facilities. BMLL covers major exchange groups such as Nasdaq, NYSE, the London Stock Exchange, Xetra, the Hong Kong Exchange, CME, Eurex and ICE. Every BMLL venue is delivered in the same harmonised format, so venues can be compared directly. The full list of venues, MIC codes, history start dates and available datasets is on the BMLL Market Data Coverage page. 
 

Which asset classes does BMLL cover?

BMLL covers five asset classes: equities, ETFs, futures, options and prediction markets. All BMLL asset classes use the same harmonised data model. 
Using one data model across asset classes means BMLL users can analyse equities, futures and options together without converting between formats. 

Does BMLL have historical equities data?

Yes. BMLL provides historical equities data across primary exchanges, MTFs, ATSs and trade reporting facilities worldwide, with full-depth Level 3 available on many venues. Depth varies by venue (Level 1, 2 or 3); see the Market Data Coverage page. BMLL equities analytics include EBBO and NBBO reconstruction, displayed depth, queue behaviour and trade-through analysis. BMLL Trades Plus, BMLL Daily Classified Trades and the full BMLL Daily Analytics library are available across most of BMLL’s equity universe. BMLL equities data can be limited to a client’s chosen instruments and venues. 

Does BMLL have historical ETF data?

Yes. BMLL provides Level 3 historical ETF data across listings and across lit and dark venues. Primary and secondary market trading are visible down to individual messages. ETF issuers use BMLL to assess the activity of authorised participants and market makers, to compare performance across venues, and to support pricing of new issuance. Buy-side desks use BMLL ETF data to assess sources of liquidity, route orders and benchmark execution. 
 

Does BMLL have historical futures data?

Yes. BMLL provides harmonised Level 3 historical futures data from CME, Eurex and ICE, plus venues including ASX 24 and B3, in one format.  BMLL futures data covers equity indices, fixed income and government bonds, short-term interest rates, FX, commodities and digital assets. BMLL futures data supports comparisons across venues, including each contract’s market share, spreads between related products and how trading moves between venues. BMLL handles contract rolls and builds continuous contract series. 
 

Does BMLL have historical options data?

Yes. BMLL provides historical US options data from OPRA, the US options consolidated tape. BMLL’s OPRA dataset is one of the largest historical options datasets available, cleaned and normalised across the full OPRA history. OPRA covers last sale prices and the best bid and offer from every SEC-approved options exchange. BMLL provides the raw OPRA messages and ready-made views, including the national best bid and offer (NBBO) consolidated to one-second intervals, trades and trade statistics. 
 

Does BMLL have prediction market data?

Yes. BMLL provides historical data for regulated prediction markets, including CME Event Contracts and Kalshi. Kalshi’s order-to-book data is normalised into the same unified schema used for CME Event Contracts, so both can be analysed together. Data fidelity varies by venue; see the Market Data Coverage page. CME Event Contracts are binary options on equity indices, commodities, FX, crypto and sports. Their prices act as a real-time market view of how likely an outcome is. Because BMLL uses its standard options format, prediction market data fits into existing futures and options workflows. 

How far back does BMLL historical data go?

How far back BMLL history goes depends on the venue. For example, BMLL data for Nasdaq starts in October 2015, NYSE in November 2015, the London Stock Exchange in March 2017, the Hong Kong Exchange in March 2018 and Xetra in February 2019. Start dates vary by dataset as well as by venue. BMLL keeps original source data indefinitely, so its history can always be rebuilt from source. Start dates for every BMLL venue and dataset are listed on the BMLL Market Data Coverage page. 

How often is BMLL data updated?

BMLL data is updated every trading day and delivered T+1. BMLL collects each venue’s data at the end of the trading day and processes it overnight, so it is ready the next day. 
T+1 delivery gives research, TCA and surveillance teams a complete, clean record of the previous session without running their own overnight data processing. 

How precise are BMLL timestamps?

BMLL timestamps are precise to the nanosecond. BMLL keeps every available timestamp, including packet and exchange timestamps, in both normalised and raw form, along with exchange sequence numbers. Nanosecond timestamps and sequence numbers let BMLL users replay the market in exact order, line up events across venues and link their own orders to market events. 

How can I access BMLL data?

BMLL data can be accessed in five ways: a REST API with a Python SDK, Snowflake, Databricks, automated cloud transfer to AWS, Azure or Google Cloud, and SFTP. BMLL data can also be used in BMLL Data Lab or BMLL Vantage. BMLL data is also listed on Snowflake Marketplace and Nasdaq Data Link, and BMLL Data Lab is available on AWS Marketplace. 

Is BMLL data available in Snowflake?

Yes. BMLL data is delivered through Snowflake native data sharing, so you can query it without making a copy. BMLL historical Level 1, Level 2, trades and ready-made analytics can be queried straight away within an existing Snowflake environment. Snowflake delivery means no data pipelines, file transfers or infrastructure to manage. BMLL data can be joined with your own trading records, execution data and reference data using standard SQL. Sample BMLL datasets are available on Snowflake Marketplace. 

Is BMLL data available in Databricks?

Yes. BMLL Data Feed works directly with Databricks Delta Lake, so the data can be queried immediately without building a pipeline. BMLL Databricks delivery supports Unity Catalog for data governance and is designed for quantitative research and backtesting. 

Can BMLL deliver data to AWS, Azure or Google Cloud?

Yes. BMLL Cloud Transfer automatically delivers BMLL data to AWS, Azure and Google Cloud in Parquet format, a fast columnar file format. BMLL Cloud Transfer can copy data across regions and between cloud providers. It can also trigger your downstream workflows automatically when new data arrives. 

Does BMLL have an API?

Yes. BMLL provides a REST API for on-demand access to historical data, with flexible query options and a Python SDK. BMLL also offers SFTP delivery (push or pull) with SSH key authentication, scheduled delivery, custom folder structures, and compressed, encrypted files. 

Do I need exchange licences to use BMLL data?

No. BMLL is Vendor of Record and holds and manages the exchange licences, so clients don’t need to negotiate with venues themselves. A small number of venues and datasets need additional approval; your BMLL contact can confirm for the venues you need. This removes one of the slowest and most expensive steps in sourcing historical market data. 

Can I see BMLL sample data or try BMLL before buying?

Yes. BMLL offers a free trial, downloadable sample files for each venue on the BMLL Market Data Coverage page, and sample datasets on Snowflake Marketplace. 
To start a BMLL trial or ask for specific samples, use the Start trial page or contact info@bmlltech.com. 

How is BMLL data used for transaction cost analysis (TCA)?

BMLL data lets execution desks reconstruct exactly what the market looked like around every trade, down to the nanosecond. Teams can then measure slippage, market impact and adverse selection accurately. Standard post-trade TCA often relies on best-bid-and-offer data or delayed consolidated data. BMLL’s full order book data adds each order’s position in the queue, fill analysis, comparison of venues and liquidity pools, independent assessment of brokers, and consistent analysis from before the trade to after it. BMLL Trades Plus adds ready-made markouts and the prevailing best bid and offer to every trade, which speeds up TCA and best execution reporting. 

How do quants use BMLL for research and backtesting?

Quant teams use BMLL’s harmonised, nanosecond-precise order book data to research new trading strategies, backtest them and model execution across many market conditions, without building their own data systems. Cleaning and joining up fragmented data from different venues usually takes up most of a researcher’s time. With BMLL, that work is already done, so researchers spend their time on research rather than data preparation. 

How is BMLL used for smart order routing and algorithm development?

 Smart order routing (SOR) and algorithm teams use BMLL historical data, normalised across many markets, to spot trends, test routing logic and backtest SOR models in a scalable, managed environment. Kepler Cheuvreux, for example, uses BMLL Level 3 data in the smart order router for its KCx execution platform. Berenberg uses BMLL to run A/B tests on its algorithms over five years of market and trade data. 

Can BMLL data be used for market surveillance?

Yes. Surveillance and compliance teams use BMLL to reconstruct full order books to the nanosecond. That lets them investigate manipulation such as spoofing and layering, which best-bid-and-offer data alone cannot show. BMLL holds and processes the very large datasets involved, so teams can focus on spotting unusual activity, keeping full audit trails, analysing trading behaviour and building evidence. 

How does BMLL support market structure and liquidity analysis?

BMLL provides harmonised historical data and ready-made analytics that analysts use to measure liquidity, compare venues and assess how rules and products affect fragmented markets. 
BMLL’s 22-value trade classification separates lit, dark, periodic auction, OTC and systematic internaliser liquidity, so analysts can see where liquidity really sits. 

How can BMLL help analyse closing auctions and index rebalances?

BMLL lets traders analyse extra volume and spreads in closing auctions, follow how liquidity moves from continuous trading into auctions, study past index rebalances and compare auction performance across exchanges. BMLL Vantage provides these auction analytics without coding. BMLL Daily Analytics provides auction metrics that can be queried in code. 

How do exchanges use BMLL?

Exchanges and trading venues use BMLL to compare their market quality with other venues, monitor market makers, test new products and show members independent evidence of execution quality. Aquis Exchange, for example, uses BMLL Level 3 data to show that it is regularly among the top two European exchanges for being at the European best bid and offer (EBBO). SIX uses BMLL to assess the performance of its liquidity providers. 

How do asset managers use BMLL?

Asset managers use BMLL’s harmonised historical data and analytics to assess execution quality, compare brokers, monitor liquidity and strengthen their best execution and governance processes. 
Because BMLL is independent, asset managers can assess brokers using a neutral view of market activity across venues. 

How do hedge funds and systematic trading firms use BMLL?

Hedge funds and systematic trading firms use BMLL’s granular Level 3 order book data to research and back-test strategies, build alpha signals and model market impact.
Because BMLL normalises data across venues, quant teams spend less time cleaning data and more time testing ideas.

How do investment banks and brokers use BMLL?

Investment banks and brokers use BMLL to measure execution performance, run transaction cost analysis (TCA) and spot anomalies and behavioural patterns across fragmented venues.
Because BMLL is independent, brokers can benchmark their execution against a neutral view of the market.

How do commodity trading advisors (CTAs) use BMLL?

CTAs use BMLL’s historical futures data to analyse liquidity and execution costs and to back-test systematic strategies across contracts and venues.

How do quant researchers use BMLL?

Quant researchers use BMLL’s normalised historical order book data and pre-computed analytics to test hypotheses, build features and study market microstructure. They can query it directly in code, without sourcing and cleaning raw feeds.

How do SEC Rule 605 filers use BMLL?

Firms that publish Rule 605 execution quality reports use BMLL’s detailed public market data to compute the required metrics and benchmark their results against the wider market.

How do regulators use BMLL?

Regulators use BMLL to monitor market quality, assess the real-world impact of rule changes, investigate complaints with full order book detail, and track market fragmentation, all without building their own surveillance systems. Because BMLL is an independent third party, regulators avoid relying on data supplied by the firms they regulate. 

Can BMLL data be used to train AI and machine learning models?

Yes. BMLL data is AI-ready. Clients use BMLL’s harmonised Level 3 data to train AI and machine learning models, because it is consistent across venues and has nanosecond-precise detail.
BMLL was founded in a machine learning laboratory at the University of Cambridge. BMLL Data Lab offers GPU computing for model training on large historical datasets. 

What support does BMLL provide? 

BMLL support is provided by its team of client-facing data scientists, who are experienced quants and researchers. All BMLL products are fully documented, and BMLL Data Lab includes pre-configured notebooks. BMLL Data Lab Enterprise clients receive dedicated support.